A Regression Equation for Determining the Dimensionality of Data.

Area of Science:

• Statistics
• Data Analysis
• Psychometrics

Background:

• Parallel analysis is a common method for determining data dimensionality using eigenvalues.
• Existing methods rely on regression equations to estimate expected eigenvalues, which can be complex.
• Accessibility for researchers is a key consideration in developing new analytical tools.

Purpose of the Study:

• To introduce a novel regression equation for estimating the mean value of eigenvalues in parallel analysis.
• To enhance the accessibility and usability of parallel analysis for researchers.
• To present a more performant alternative to existing eigenvalue estimation methods.

Main Methods:

• A new regression equation was developed to estimate the mean eigenvalue.
• The performance of the new equation was compared against previously published regression equations.
• The study focused on estimating eigenvalues for sample correlation matrices assuming an identity population correlation matrix.

Main Results:

• The proposed regression equation demonstrated favorable performance compared to existing equations.
• The new technique simplifies parallel analysis by eliminating the need for coefficient tables.
• The method provides a more accessible approach to dimensionality assessment.

Conclusions:

• The novel regression equation offers an improved and more accessible method for parallel analysis.
• This advancement facilitates more accurate and efficient data dimensionality determination.
• The technique reduces complexity, making advanced statistical methods more readily available to researchers.